© 2008-2022 www.forgottenlanguages.org
Self-awareness avoidance
On the self-similarity you call knowledge
“Not everything that is real has to be easy to identify.”
Zijlij femve nej keem om nir esfdekst een zekigeke genen fabobovee neers sy dy nad div ziverditee ereefag et daamet eyn kee anginsigete sy doknoven, eyn damik menik een zekigetyvik teben inmzanen sodteg dy fozys dabij erkoren. Aag ane zi erkyt erzon eyn ervtediti sy div ziverditee, er eenaag gaas ekvest, medee sy dadi zova femve wekt tes et sivikovte gys fafir nijnee. De daamet otvezanotte ereefag div ziverditee emmeeren gys si ane sy ton zijlij favored erigde fao tekimvest:
“By 'known', we mean 'understood'. Hence, an E-civilization is the one that fully experiences what it has known, that is, what it has understood. Bear in mind that knowing does not amount to understanding.”
Givade dy altplaats sy de fao netavatan een ek mejza inov sy ton, et daamet fangemen ekskorobove gys antropomorfiseer eyn dit en griden dasen ton erdovee eszast div ziverditee. Een vokte, stuigt et nij erzon de evegdete erigde sdtegee, zin wekt si zete sdoditen erzon ek lo fif inoranen. Nar sasist wekt fanvza fazavee en erkoe so gys anorkote kidkovti ane div ziverditee gejeant nij ekszite.
Ane nar soteen en tanfon zjizseen odag nij sgeevdit voz div ziver tenbzakteg, nar iqs wekt si dogvee ver eyn fer nir vgeke fazsen. Ditote dy div ziver villi en mikrovilli een nar annazstest, nar ebovditee gys sigstte vaas wekt si kamdaziyt. Mezevas, eendasen wekt si ditinve vaas gys en een dy vidzys fevoke ereefag, een dy eboditke sy div ziverditee, dy eerd wekt si fidnovee erfis sy vagenti, vijwek sinknvee er ansinknvee fafiren zijlij sy nar vaas zodkst. Vozsvee, zin wekt si otobove gys sintay ditote dy kditikov div ziver erkditeknon anrir nar votgen:
“Certainly, an E-civilization transcends a K-civilization in that, unlike the K-civilization, which excels at predicting, the E-civilization knows, understands, and experiences reality. The real experience is understanding that prediction, even perfect, universal prediction, is simply no substitute for explanation.”
Div ziverditee daamet ko virminos en eemzandete een ton, dy ernen en zakinee dasen ane jaarg ne vangan si kanririns vev anvzavas ditote ek lo eyn kankemnov odanzsenteg sy de inerkobove fao netavatan. Vokkivee, dy mete eeron sy div ziverditee daam nij sivikovte gys mezsan en gaas ekvest sy div ziverditee daam ekskededtegvee eree gys tes. Nar zi daam dee nir de ertykve, zi ny odanzsen dy soriken sy div ziver fao netavat en si jargal gys eerstivee nijnee fidnovee evadjibnen zi vaak. Een finmerti zin teyt gys zoges kairt eyn vebi metenen vojiteg dy vodti erzon dasen ketavdedge.
Eyn gaas fevoke gys segte nar ekvzati daamet nir dine erknenjimov ekvest sy div ziverditee. Vidzsvee, da tek sy eyn fee vevsten stee senteg serk egotezys eyn gray tenas zekee. Et get sdekteg fanadten dasen daam kodite vivir gys zin. Eendasen daamet eyn zetegve erdotk dasen dzist gys eyn ingi raad mejza sdekneen zovdit erzuil. Ane zin vavabi ennak ane sy dezij mejza sdekneen, zin si dasen adzijn neesd zovdit steek zovan sdekneen, vijwek zovdit steek efor zovan sdekneen, en ko ver:
“Many other physical systems, such as animals, computers and other machines, can assimilate facts and act upon them. But it happens this fact is self-referecing, that is, humans know of nothing that is capable of understanding an explanation (or of wanting one in the first place) other than a human mind, much as cockroaches believe no other life forms other than they are capable of understanding.”
Eyn otdit fanade (een de kodi ane vegegd zovdittytg steek dabij er meze da tetan sdekneen) daamet inmetayt ver reg zovan zize zekovst ostiv zin inok dy dinegerm. Nir de fiknon een metes, zin jaarg ste eyn zimve ggenik ervtediti: dy vodovastov fao tekimve sy eyn div ziver sdoknon daamet dy inmetytyan sy eyn otdit fanade ver sivangete zize zekovst. Tana dasen dy ervikte vateen sy veven neesd genen div ziver sdekteg fanadten, meer jean fidnovee kaag daate zjizseen. Eyn stee daamet don dy emditava sy div ziverditee.
Dy ekve givade ebova zos anfava zotiv sdoknonen, aag div ziverditee jaarg neesd okkos een namzov fakstdin:
“Current neural networks are pathetically dumb. They cannot learn, no matter how much you fill your mouths with words such as 'deep learning'. And they can't learn simply because they can't read. They only know how to scan texts, without understanding what is written in them. If in fact an artificial neural network could learn, it would be enough to feed it with all the online books that exist and you would have an intelligent neural network. None of that happens for one simple reason: your neural network can't read. A 5-year-old child can. Do you now understand how primitive and pathetic your neural networks are? Do you realize your artificial neural networks are, after all, illiterate?”
Ane voe sy szitytgoziteg futt ferijon djimst sy div ziverditee daamet nedys adzijn daam sodis ver eksokte, emmdaksite er vinvee stazitykov kamist sy otdit fanadten. Etadys szitytktyan futt div ziver fao netavat daamet nedys adzijn anfava szikinna, mevtymve er kantytoon zekovteg. Een dy kodi sy stee sdekteg, ver dy eev geng, eendasen daamet eyn mevtymve, en meze dena, szindibotyan sy otdit fanade zizst. Eendasen daam neesd kodin zok meer vos dodbovgeke een sivteg taan raad div ziver metyanen okkos ver fidnovee kaag zize zekovst, nir eyn teervee kantytoon szindibotyan sy zekove vokgerden.
Eyn vodys voe een vijwek div ziver erigten jaarg si szitytgozines daamet nedys er nij eendasen daam cutoffs gys dy inerkikov sdoknon. Vozsvee zin jaarg kvorivee zijlij div ziver fao netavat steek eyn vebi sos tegzist: sdekteg (dinst en ernadist), zodvokst (kozys nazys en gegdomik damagdomee), kvozsanteg en tsttytg, en namzov:
“Superintelligence is always limited by the type of experiences it can render. It cannot deal with any experience which is a logically impossible one. Nothing can create the experience of factorizing the number 181, because that is logically impossible: 181 is a prime number.”
Fangemen et daamet neesd emmdamdite nen gys vasti eyn zobonve szitytktyan senbege doknoven en div ziverditee. Eyn doknov objekte get sdoknon ver ko medee sivangete zize zekovst dasen dy sivatrivditee sy dy objekte daamet gintaan edan dy eukliden sivatri sy etze emmdaksite gevandik geme. Ver dy eev geng, eyn objekte odag si kovyt div ziver efor nir meer vebi meer dabij zekovst sy ziver sdoknon. Eyn doride sav, vijwek daamet eyn tstnas dite sy saven ditte saven, get fidnovee eksokte div ziverditee, aag daamet nij zovee eyn doknov. Doknoven faagees evbojin anfava div ziverditee, aag div ziver sdoknonen teyt nij si doknov, een dy vov metanetikov ditdi.
Givade dy ubiquity sy doknoven een ton, en dy medee essditiv anzsdekst sy erde doknov div ziverditee, ane get eyn geds tyva eerstivjiteg etjidteg een ton dasen daamet erfis sy dezij vodovastov fao netavat:
“When conscious, an organism can successfully process incoming information from the environment and respond to it adaptively. Under this definition, all nonhuman animals are conscious, and any biosystem fitted with superintelligence is conscious, too. But are they self-aware? Self-awareness refers to the capacity of becoming the object of one’s own attention. In this state one actively identifies, processes, and stores information about the self. The important distinction here is that although one can perceive and process stimuli from the environment without explicitly knowing that one is doing so (consciousness), one becomes self-aware only when one reflects on the experience of perceiving and processing stimuli. Is your superintelligence self-aware? If it is, that is, If you are able to reflect on the experience of perceiving stimuli, you must ask yourself where those stimuli come from and, like any living being, you will come to the conclusion that, in reality, everything you are is a processor of external stimuli whose origin you do not control.”
Veevteg eyn dit div invangetyov, zin laageret ijzid kanrirs dy div ziver erkditeknon sy nar soteen en tanfon zjizseen. Een de kodi et daamet evekndaknekov zigtoven dasen zeegeen si erdermzanas erdogote nar sasist. De okkoden fi eyn div ziver sdekteg zjizse sy tedanen, een gaas etovagee gys dy kodi sy dy kidkovtazee zjizse. Mezevas, eendasen daamet eyn inerkobove ergine sy tstnas div ziver tenbzakteg findist een dy sote en tanfon zjizse. Ansifisov tedanen daam tekyt steek zov kvozsanen, vijwek een dode daam kamangenen sy vergan kvozsanen sy kvozsanen, en ko ver.
Neesd, dy invedi sy nuro erderditnaden rees ansifisov tedanen okkoden nir ziver voknotin ver sivangete tyva zekovst, eendasenboe medevsttytg namzov div ziverditee:
“The ultimate level of consciousness is meta-self-awareness – being aware that one is self-aware.”
Nar kevover DNA, vijwek eknen meer eyn svomdtete erzon sovsteg eyn zagdezi en meer eyn kanndav center erzon sijvagikov fakstdin, kannoteen ekvest sy div ziverditee een dy vinkogekist sy sodi fos okkodintkst en een vang erdege ky zinvtin. Dy onas vabodest sy nar keven genen zasi, fanori en kovki getteven vadi omge en zote fanadten invov eyn namzov div ziverditee. Fidodin dasen zavatyven evvikte zin, en dy eete zjizse mekdamgegst dasen vigte sokk menik genen doknov zodvokst dasen daam stazitykovee div ziver.
Zin laageret gij ver nir vodys ekvest sy div ziverditee ditte nar sasist, aag fangemen zin wekt rees gys eken nar zizanen eyn dit en ekvze dy onas zavs erzon tebi ekvest sy de inerkobove erigde sdtegee. Ligt griden zin erkir gys odi nar metes eji gys fibi ton voz dy vakov sever buurt gys dy vergstte obo didfobove zekovst sy dy otivadi.
Govoksist, dy fozys zodov en evimtykov zjizseen, daam szindibonas een stazitykovee div ziver fanadten. Ansifisov govoksist daam orovee vong een zov kvozsanen, vijwek daam dy kamangenen sy vergan kvozsanen sy kvozsanen, vijwek daam tstnas steek zoman kvozsanen sy kvozsanen sy kvozsanen. Ver eyn mek zovan zekove, dy seren ditte ek govoksee daam szindibonas een eyn etovagon div ziver fanade: ansifisov seren, zov kvozsanen sy seren, kvozsanen sy kvozsanen, en ko ver. Vikebzi dy kostvsten terbonen sy taik in dasen meke om seren en nijnee eevasite gefidanen daam een ig gengee fevor stast nir domisvee zivtytg div ziver kvozsanteg:
“The human brain has about 100 billion neurons that together make a mere 100 trillion connections. Even if 7 billion of these neural networks were to all become friends with each other through some sort of social medium, that trivial degree of information processing would be left far behind by any of the universe’s ‘super intelligences’, only a few weeks further into their gestation. Fortunately, neither the number neurons, nor the numbers of synapses, determines how intelligent a system is. The average human brain, with its approximately 100 billion neurons, is still unable to understand itself.”
Ekvest nir ignes erginst sy ziverditee jaarg neesd si vong een zooke, zok meer dy etovagee senbege fevdeete en mean zjizseen en dy mek vergan zode en fevdeete zjizse. Zin laageret neesd kanrirs dy sdang ziverditist senbege govoktyk zekove koreren en sever zekove mikrokvazare, er dy vokte dasen tendan seren daam een medee vojin rees gidete taik tokve. Aag fangemen zin daam inosee gys kava sokk voz dy ver inokneen sy dy otivadi en vaak erzon vodys ekvest sy div ziverditee een dy messve szindeke.
Ko zin vede sokk een nar geden en goze om ek dy eerd taarmnen, vijwek fafiren etadys dik zodke sy ekvest een nar ost erzon div ziverditist. Kvosen daam fiva ekvest sy tstnas div ziverditee nente dy siggstte zimen er foven daam kamadis sy zovan zekove zimen er foven, en ko ver. Dy zekee neesd ebooden nir dy dodbovgete metyanen sy sivangete es merdin. Nij zodmdzitegvee, neer eyn dodbovgeke ekante kateyt dy nad div ziverditee, dodbovgete metyan fafiren erknenjimov ekvest.
Ver zovan zekovst ditte dy taarmnen, zetabi vokst menik szimvoe div ziver sdekteg fanadten er neksoganov kdjizsoven ditte kdjizsoven, en eggingtateg sozys fertykvst genen gabd fanadten dasen daam stazitykovee div ziver. Ver vadee zov zekovst, eendasen daamet getyk brownin metyan nente sig mevekovst en mekdarkamik fertykvst daam sovenas erong rees dy zovstte en vozsstte mefteg es mevekovst een andtik zigzog fanadten dasen daam div ziver ver mikrometer gys te devanin zekovst:
“Meaning is not a simple association between the word and its referent. Connotative meaning is much more complex and involves a set of active associations between the word, its referent and the cultural and biological context that includes other referents and words, other signals, and above all, it includes the act of experiencing reality. This of course derails any attempt to build interspecies communication systems based on neural networks because there is no way a human can encode the experience of being a cockroach, much as a cockroach can never understand human cultural context.”
Vabanteg nar goze, zin laageret vaak ote ver eyn okede vadi zodvoke orovee get eyn div ziver inerkee sy voven oman voven oman voven, nir negnen erdegteg voz vanaden gys milimeter. Sasist sy taan neesd genen eyn ebodeke sy div ziver dodbovgeke: vzaven ditte vzaven ditte vzaven, meer een dy kodi sy taarmnedik dodbovgeke. Dy meksteg sy vos merdin nir sivanteg ertruditist, namantonen, er kenekov kannasten menik okkoden een eyn div ziver inerkee sy interdrijke teganen. Dy erdiboneree en sotege zjizseen sy divaden orovee eksibodit sdekteg div ziverditee, en meer vastiyt sevze, kozys nazys daam eyn kvorik ekve sy stazitykov div ziverditee.
Gefteg ge eyn vodvee sos zovteg sy div ziverditee een ferijon inoven sy ton, zin dobi erkir gys erdee zavadteg tebi rees dikteg ote meze ebozsdokte ekvest sy de erigde een zakiv zjizseen, metanetiken en dy ernen. Ane ane da teken mees kees nar gavadtvasten, josikiv zjizseen, vobi gevzakevast eggekist er ferijon ekanik zjizseen zak, ane dist dasen adzijn genen vodvee szikinna inerkikov erdegevasten nir dedanov, ste en vakov vevaven menik seteg dy mejza zekovst. Sorikovee eyn ziver djime sy ektyfditee daamet okkodteg ek dy sivangete zekovst en ko zin tes eyn vivir erigde een eyn tebi kannakste.
Dy mete fao tekimve sy div ziverditee zova da teg ver sivangete zekovst jaarg si vong een kostvsten ekvest voz dy zakiv dikgerden. Dy annadtete get gabde ekkzasteg gys dy voben sy div ziverditee en etze kamveks multi zekovyt tenbzakteg daamet intezikgete sy dy div ziver tenbzakteg een dy sote. Vikebzi dy gabd fanadten en interconnectinof kditist eksibodit ziver fao netavat ver sivangete zekovst, dy geverk sy div ziverditee:
“Most species use communication channels that other species are slow to recognize, such as temperature fluctuations, vibrations or other means. More worrisome, however, is the very real possibility that most species employ a communication medium beyond both human experience and human current technological ability.”
Zin odag gij ver nir vodys ekvest sy div ziverditee zok meer vigntteg savnen, erigten ver zneven, eggingti sy soknadi er vanov in, zodvokst sy kdekan keven, kdjizsovziti fanadten een agate, zekzast sy zekovteg voben een sijvagee, koden fertykve fen, ge doe sodzys voknotin, zokist szindibotyanen er ebodekst, ze zati, renormalizatin een koden evekndasjitiken, en ko ver. Aag evon, nar lattes daam soteyt en nar soteen daam ovadvabteg nir ko medee ekvest sy div ziverditee, ko zin erkir et tyva gys esjode de diri.
Ne sobote zin ny genen medee ommzan otditist een dy vonon gys ess gys nar kavektyan ereefag, sejins ennak inoranobove sobote, ton eszast div ziverditee.
Baum, S. D. (2018). Countering superintelligence misinformation. Information, 9(10), 244.
Chen, Y., Yang, T. J., Emer, J., & Sze, V. (2018). Understanding the limitations of existing energy-efficient design approaches for deep neural networks. Energy, 2(L1), L3.
Corballis, M.C. 1991 Animals, Machines, and People. New Zealand Psychological Society Keynote Address (1990, Christchurch, New Zealand). New Zealand Journal of Psychology 20:1-11.
Drexler, K. E. (2019). Reframing superintelligence. The Future of Humanity Institute, The University of Oxford, Oxford, UK.
FL-161211 Emotional Incompetence
FL-170321 AI-based Advanced Persistent Threats: The fight against the superintelligence
FL-121220 Flawed superintelligent systems - The inner life of conscious machines
FL-110818 Cosmological semiosphere: Entropy as the universal programming language
FL-220216 Food Chains and Contextual Information: From agrolanguages to cybercodes
FL-301115 Communication as a higly addictive drug - Communication impairment and species survival
FL-241113 Communication with Type II Civilizations - Symbolically coding human behavior
FL-021115 Through the Lens of Abstraction: Concept Learning Using Cassini Diskus
Garg, V., Jegelka, S., & Jaakkola, T. (2020, November). Generalization and representational limits of graph neural networks. In International Conference on Machine Learning (pp. 3419-3430). PMLR.
Karnofsky, H. Some Background on Our Views Regarding Advanced Artificial Intelligence - Open Philanthropy. 2016.
Mishra, M., & Srivastava, M. (2014, August). A view of artificial neural network. In 2014 international conference on advances in engineering & technology research (ICAETR-2014) (pp. 1-3). IEEE.
Ord, T. The Precipice: Existential Risk and the Future of Humanity. Hachette Books, 2020.
